back to: [Main Page]
[Table of Contents]
[12) Soil Orders]
back to: [Main Page]
[Table of Contents]
[12) Soil Orders]
Some definitions are given below which deal with soil mapping.
Pedon: A three-dimensional body of soil with lateral dimensions large enough to permit the study of horizon shapes and relations. Its area ranges from 1 to 10 m2. Where horizons are intermittent or cyclic, and recur at linear intervals of 2 to 7 m, the pedon includes one-half of the cycle. Where the cycle is <2 m, or all horizons are continuous and of uniform thickness, the pedon has an area of approximately 1 m2. If the horizons are cyclic, but recur at intervals >7 m, the pedon reverts to the l m2 size, and more than one soil will usually be represented in each cycle.
Polypedon: A group of contiguous similar pedons. The limits of a polypedon are reached at a place where there is no soil or where the pedons have characteristics that differ significantly.
Soil map unit: (i) A conceptual group of one to many delineations identified by the same name in a soil survey that represent similar landscape areas comprised of either: (1) the same kind of component soil, plus inclusions, or (2) two or more kinds of component soils, plus inclusions, or (3) component soils and miscellaneous area, plus inclusions, or (4) two or more kinds of component soils that may or may not occur together in various delineations but all have similar, special use and management, plus inclusions, or (5) a miscellaneous area and included soils. (ii) A loose synonym for a delineation.
A map unit is a collection of areas defined and named the same in terms of their soil components or miscellaneous areas or both. Each map unit differs in some respect from all others in a survey area and is uniquely identified on a soil map. Each individual area on the map is a delineation. Map units consist of one or more components. An individual component of a map unit represents the collection of polypedons or parts of polypedons that are members of the taxon or a kind of miscellaneous area. A delineation of a map unit generally contains the dominant components in the map unit name, but it may not always contain a representative of each kind of inclusion. A dominant component is represented in a delineation by a part of a polypedon, a complete polypedon, or several polypedons. A part of a polypedon is represented when the phase criteria, such as a slope, requires that a polypedon be divided. A complete polypedon is present when there are no phase criteria that require the subdivision of the polypedon or the features exhibited by the individual polypedon do not cross the limits of the phase. Several polypedons of a component may be represented if the map unit consists of two or more dominant components and the pattern is such that at least one component is not continuous but occurs as an isolated body or polypedon. Similarly, each inclusion in a delineation is represented by a part of a polypedon, a complete polypedon, or several polypedons. Their extent, however, is small relative to the extent of the dominant component(s). Soil boundaries can seldom be shown with complete accuracy on soil maps, hence parts and pieces of adjacent polypedons are inadvertently included or excluded from delineations.
Soil association: A kind of map unit used in soil surveys comprised of delineations, each of which shows the size, shape, and location of a landscape unit composed of two or more kinds of component soils or component soils and miscellaneous areas, plus allowable inclusions in either case. The individual bodies of component soils and miscellaneous areas are large enough to be delineated at the scale of 1:24,000. Several to numerous bodies of each kind of component soil or miscellaneous area are apt to occur in each delineation and they occur in a fairly repetitive and describable pattern.
Soil map: A map showing the distribution of soils or other soil map units in relation to the prominent physical and cultural features of the earth's surface. The following kinds of soil maps are recognized in the USA: (i) soil map, detailed - A soil map on which the boundaries are shown between all soils that are significant to potential use as field management systems. The scale of the map will depend upon the purpose to be served, the intensity of land use, the pattern of soils, and the scale of the other cartographic materials available. Traverses are usually made at 400-m, or more frequent, intervals. Commonly a scale of 10 cm = 1609 m is now used for field mapping in the USA. (ii) soil map, detailed reconnaissance - A reconnaissance map on which some areas or features are shown in greater detail than usual, or than others. (iii) soil map, generalized - A small-scale soil map which shows the general distribution of soils within a large area and thus in less detail than on a detailed soil map. Generalized soil maps may vary from soil association maps of a county, on a scale of 1 cm = 633 m, to maps of larger regions showing associations dominated by one or more great soil groups. (iv) soil map, reconnaissance - A map showing the distribution of soils over a large area as determined by traversing the area at intervals varying from about 800 m to several kilometers. The units shown are soil associations. Such a map is usually made only for exploratory purposes to outline areas of soil suitable for more intensive development. The scale is usually much smaller than for detailed soil maps. (v) soil map, schematic - A soil map compiled from scant knowledge of the soils of new and undeveloped regions by the application of available information about the soil-formation factors of the area. Usually on a small scale ( 1:1 000 000 or smaller).
The basic distinction between soil mapping units and soil taxa is that the latter is an abstract concept in that it is a grouping according to specific ranges of soil properties for purposes of scientific categorization, whereas a soil mapping unit is a cartographic representation on a map of the polypedons as they actually oocur in the field.
In chapter 11.1. Soil Classification the categories used in the U.S. Soil Taxonomy were listed ranging from orders, suborders, great groups, subgroups, families, to series. Additionally, the terms consociations, taxadjuncts, and variants are used to describe inclusions of areas of small differences from the main soil units. They help to define the degree of geographic purity of soil mapping units.
Consociations: Mapped areas dominated by a single soil taxon and similar soils. At least half of the pedons on each delineation of a consociation are of the same soil components that supply the name for the map unit. Much of the remainder of the mapping unit consists of soils so similar to the named soil that major use and management interpretations are not significantly different. Generally, the total area of dissimilar inclusions of other components in a map unit does not exceed 15 to 25 %. A single component of dissimilar inclusions generally does not exceed 10 % if very contrasting.
Taxadjuncts: (i) Polypedons with properties outside the range of any recognized soil series and exceeding the higher category class limits by one or more differentiating characteristics of the series. (ii) A soil that is correlated as a recognized, existing soil series for the purpose of expediency. They are so like the soils of the defined series in morphology, composition, and behavior that little or nothing is gained by adding a new series.
Variants: A soil with characteristics outside the limits of any known soil series and which is less than 2000 acres in extent is classified as a variant.
Soil surveys produced by the United States National Cooperative Soil Survey are described as being at least 85 % pure (Soil Survey Division Staff, 1993), although field checks suggest that the figure may be lower.
Reference:
Soil Survey Division Staff. 1993. Soil Survey Manual. Handbook 18, US Govt. Printing Office, Washington, DC.
back to: [Main Page]
[Table of Contente]
[12) Soil Orders]
The purpose of soil classification is to reduce a complex system of varying soil characteristics into explicitly defined classes. Soils occur as a continuum in nature, however, crisp classes are used to distinguish soil map units, which differ in one or more characteristics from each other. These soil map units are our best approximations of what we perceive to be truths. Soil mapping scales range from coarse (small) to fine (large) scale (Table 13.1.1 and 13.1.2.).
Table 13.1.1. Soil mapping scales (Buol et al. 1997).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Table 13.1.2. Information of scales of soil maps and map units representing them (Buol et al., 1997).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Some users of soil surveys need very specific and detailed information about soils. For these potential users, the information needed is about the nature of soil areas of a few hectares or less. Other users may need only a broader soil information such as areas of thousands of hectares each. Therefore, different levels of detail are provided in the soil survey maps. These sizes and levels of detail are arranged in classes of soil surveys called 'orders of soil survey' (Soil Survey Division Staff, 1993). These orders differ in kind of map units reflected in the soil survey legend as consociations, complexes, and associations.
Table 13.1.3. Order of soil survey (Soil Survey Staff, 1993).
|
|
|
|
First order: very intensive - experimental plots, building sites |
|
|
Second order: intensive - general agriculture; urban planning |
|
|
Third order: extensive - range land, community planning |
|
|
Fourth order: extensive - for broad land use potential and general land management |
|
|
Fifth order: very extensive |
|
References
Buol S.W., Hole F.D., McCracken R.J., and Southard R.J., 1997. Soil Genesis and Classification. Iowa State University Press, Ames, Iowa.
Soil Survey Staff. 1993. Soil Survey Manual 18, US Govt, Printing Office, Washington, DC.
back to: [Main Page]
[Table of Contents]
[12) Soil Orders]
Soil survey: (i) The systematic examination, description, classification, and mapping of soils in an area. Soil surveys are classified according to the kind and intensity of field examination. (ii) The program of the National Cooperative Soil Survey that includes developing and implementing standards for describing, classifying, mapping, writing, and publishing information about soils of a specific area.
The National Cooperative Soil Survey (NCSS) is a joint effort among the Natural Resource Conservation Service (NRCS), land-grant universities, and other state and federal agencies with an interest in the soil resource.
Soil surveys are carried out by soil scientists with good experience in soil descriptions and soil forming processes. Aerial photographs are used to determine land use pattern, drainage, and some other characteristics of the soil surface. Stereo photography is used to analyze the topographic attributes such as elevation, slope, and slope shape. Soil data are derived by sampling with augers or soil pit descriptions. For the development of a detailed soil map many data have to be collected in the field for soil classification and the delineation of the boundary for each soil map unit. Comparisons to other soil data (e.g. nearby counties) are necessary. Finally, digital orthophotos are used in the process to derive detailed soil maps. Additionally, representative soil samples are analyzed in the laboratory. Interpretations regarding the suitability of these soils for various land uses are based on detailed understanding of soil characteristics, field experience, and consultation with local landowners and other experts located in the county and state.
Soil surveyors (keen observers with experience) have the ability to integrate the soil forming factors and processes. However, the computer with appropriate software programs is now as important a tool in soil survey as the auger, spade, map board and color book.
Soil surveys were carried out for most of the United States. The Soil Survey Manual provides the major principles and practices needed for making and using soil surveys and for assembling and using data related to them. The Manual is intended primarily for use by a soil scientist engaged in the classification and mapping of soils and in the interpretation of soil surveys.
More details about soil and soil survey can be found in the Soil Survey Manual - chapter 1.
To answer this question an example is given which can be found in the Soil Survey of Columbia County, Wisonsin (1978, United States Department of Agriculture, Soil Conservation Service in cooperation with Research Div. of the College of Agriculture and Life Sciences University of Wisconsin). The contents of this soil survey comprise information about:
General soil map
associations.
Example: Plano-Griswold-Saybrook association
Well drained and moderately well drained silty soils that have a silty or loamy subsoil; underlain by sandy loam glacial till. This association is on glaciated uplands where the soils formed in loess and the underlying glacial till. The landscape is one of long low drumlins and ground moraines characterized by long slopes, swells, swales, and some broad depressions. Moderately steep, glaciated limestone ridges or higher topography parallel the larger drainageways.
This association makes up about 16 % of the county. It is about 50 % Plano soils, 14 % Griswold soils, 10 % Saybrook soils, and 26 % soils of minor extent.
Plano soils, on swells, are mostly gently sloping and are well drained or moderately well drained. Typically, the surface layer is silt loam, and the subsoil is mostly heavy silt loam. Calcareous sandy loam till is at a depth of more than 60 inches.
Griswold soils, on the crests of drumlins, are mostly gently sloping and sloping and are well drained. Typically, the surface layer is silt loam, and the subsoil is mostly sandy clay loam. Calcareous sandy loam till is at a depth of about 38 inches.
Saybrook soils, on small rises and drumlins, are mostly gently sloping and sloping and are well drained. Typically, the surface layer is silt loam, and the subsoil is silty clay loam and loam. Calcareous sandy loam till is at a depth of 38 inches.
Less extensive in this association are Channahon, Joy, Ringwood, Ripon, and Troxel soils. Channahon and Ripon soils are along limestone ridges. Joy soils are on terraces along drainageways and in depressions. Ringwood soils are on drumlins. Troxel soils are in areas that receive sediments from adjoining soils.
This association is well suited to crops. The soils have a thick surface layer high in content of organic matter, and most have high fertility and available water capacity. The main concern in management is controlling water erosion. Improving drainage is a concern in some low-lying areas. This association is used intensively for crops, mainly corn. Steeper areas are in permanent pasture.
Description
of the soils (soil series).
The soil series category is the most homogeneous category in the taxonomy used in the United States. As a class, a series is a group of soils or polypedons that have horizons similar in arrangement and in differentiating characteristics. The soils of a series have a relatively narrow range in sets of properties. The surface layer and such features as slope, stoniness, degree of erosion, and topographic position may vary unless these factors are associated with significant differences in the arrangement of horizons. Soil series are differentiated on all the differentia of the higher categories plus those additional and significant characteristics in the series control section. Some of the characteristics commonly used to differentiate series are the kind, thickness, and arrangement of horizons and their structure, color, texture, reaction, consistence, content of carbonates and other salts, content of humus, content of rock fragments, and mineralogical composition. A significant difference in any one of these can be the basis for recognizing a different series. Very rarely, however, do two soil series differ in just one of these characteristics. Most characteristics are related, and generally several change together.
Example: Plano series
|
Series |
Family |
Subgroup |
Order |
|
Plano |
Fine-silty, mixed, mesic |
Typic Argiudolls |
Mollisols |
The Plano series consists of well drained and moderately drained silty soils. These soils formed in 40 to 60 inches of silty sediment and the underlying sandy loam till. They are on silt covered till plains. Slopes are typically 0 to 8 % but range up to 12 %. In moderately well drained areas slopes are typically 0 to 4 %. The native vegetation was mixed grasses.
In a representative profile the surface layer is silt loam about 13 inches thick. The upper 9 inches is black and the lower 4 inches is very dark brown. The subsoil extends to a depth of more than 60 inches. The upper 13 inches is dark yellowish-brown, friable silt loam and heavy silt loam; the next 19 inches is dark yellowish-brown light silty clay loam; the next 6 inches is dark yellowish-brown, friable loam; the lower 9 inches is dark-brown, friable heavy sandy loam.
In some places, these soils are saturated at a depth of about 3 to 5 feet for significant periods during wet season. Permeability is moderate. Available water capacity and fertility are high. These soils are well suited to crops and are cropped intensively. Corn is the major crop. Some areas are in peas or small grain.
Representative profile of Plano silt loam, 2 to 6 % slopes, 147 feet west and 174 feet north of the southeast corner of NE1/4NW1/4 sec. 35, T. 13 N., R. 11 E.
Ap - 0 to 9 inches: black (10 YR 2/1) silt loam; moderate, fine, subangular blocky structure; friable; common fine roots; few fine and medium tubular pores; neutral; abrupt, smooth boudary.
A12 - 9 to 13 inches: very dark brown (10 YR 2/2) silt loam; moderate; medium, subangular blocky structure; friable; common fine roots; common fine and few medium tubular pores; slighly acid; clear, wavy boudary.
B1 - 13 to 17 inches: dark yellowish-brown (10 YR 3/4) silt loam; weak, medium, subangular blocky structure; very friable; common fine roots; many fine and few medium tubular pores; medium acid; clear, wavy boudary.
B21t - 17 to 26 inches: dark, yellowish-brown (10 YR 4/4) heavy silt loam; weak, fine and medium, subangular blocky structure; friable ; few fine roots; many fine and few medium tubular pores; common thin clay films on faces of peds; medium acid; clear; wavy boundary.
B22t - 26 to 45 inches: dark, yellowish-brown (10 YR 4/4) light silty clay loam; weak, medium, prismatic structure parting to moderate, fine subangular blocky; friable; few fine roots; common fine and few medium tubular pores; common thin clay films on faces of peds; medium acid; clear, wavy boundary.
II B31 - 45 to 51 inches: dark, yellowish-brown (10 YR 4/4) loam high in content of silt and very fine sand; weak, medium, subangular blocky structure; friable; few fine roots; common fine and few medium tubular pores; slighly acid; abrupt, wavy boudary.
II B32 - 51 to 60 inches or more: dark-brown (7.5 YR 4/4) heavy sandy loam; weak, medium, subangular blocky structure; friable; few fine roots; few fine and medium tubular pores; slightly acid.
The solum ranges from 44 to 77 inches in thickness. Depth to the till material ranges from 40 to 60 inches. The A horizon ranges from black (10 YR 2/1) to dark brown (10 YR 3/3) and is 9 to 20 inches thick. The B horizon is brown (7.5 YR 4/4) or dark brown (10 YR 3/3) to yellowish brown (10 YR 5/4) and ranges from 30 to 50 inches in thickness. The B2 horizon is dominantly heavy silt loam. The average clay content is 20 to 27 %. The B horizon is medium acid to neutral. The II B3 horizon is sandy loam, heavy sandy loam, sandy clay loam, loam, or clay loam. In some places the B3 horizon is mottled. The C horizon is sandy loam till.
Plano soils occur with Joy, Ringwood, Ripon, Saybrook, and Troxel soils. They have better drainage than the somewhat poorly drained Joy soils. They have a thicker solum and contain more silt and less sand above a depth of 40 inches than Ringwood and Saybrook soils. Unlike Ripon soils, they are not underlain by limestone bedrock above a depth of 40 inches. They have a thinner A horizon than Troxel soils.
Plano silt loam, 0 to 2 % slopes (PnA): Large, irregularly shaped areas of this soil are in broad slight depressions in the till plain. The profile is similar to the one described as representative of the series, but the surface layer is 2 to 7 inches thicker, and in most areas the soil is motttled at a depth of about 35 to 50 inches. Included with this soil in mapping are small areas of Joy silt loam, 0 to 4 % slopes, Troxel silt loam, 0 to 3 % slopes. Also included are a few areas where the subsoil formed entirely in silty sediment and areas where the substratum is sand and gravel outwash. This soil has no serious limitations for farming. In some places, water ponds in spring and after a rainfall. Runoff is slow. Use of the soil for septic tank filter fields is hazardous. This soil is well suited to crops. Most of the acreage is cropped to corn or peas. Capability unit I; woodland group not classified; wildlife group 4.
Plano silt loam, 2 to 6 % slopes (PnB): This soil occurs mostly as irregularly shaped areas on the foot slopes of uplands and on rises on the till plain. Most areas are 40 to 100 acres in size. This soil has the profile described as representative of the series. In the more sloping areas, 4 to 6 inches of the original surface layer have been lost through erosion, and the present surface layer is 9 to 11 inches thick. In many areas, typically where slopes are 2 to 4 %, the soil is mottled at a depth of about 35 to 59 inches. Included with this soil in mapping are samll areas of Plano silt loam, 0 to 2 % slopes, and Saybrook silt loam, 2 to 6 % slopes, eroded. Also included are a few small areas where the surface layer is loamy, areas where the substratum is sand and gravel outwash, and areas where limestone is at a depth of 50 to 60 inches. Runoff is slow. The erosion is generally slight, but where the slope is 4 to 6 %, the hazard is more serious if the soil is poorly managed. This soil is well suited to crops, but erosion is a concern in some areas. Most of the acreage is cropped to corn or vegetable crops. Capability unit II-1; woodland group not classified; wildlife group 4.
Plano silt loam, 6 to 12 % slopes, eroded (PnC2): This soil is on rises and the sides of ridges on the till plain. The areas are commonly long and less than 40 acres in size. The profile of this soil is similar to the one described as representative of the series, but the surface layer is a few inches thinner and the substratum typically is at a depth of 50 to 60 inches. Included with this soil in mapping are small areas of Ringwood silt loam, 6 to 12 % slopes, eroded, and Saybrook silt loam, 6 to 12 % slopes, eroded. In a few areas, the surface layer is 7 to 9 inches thick. Also included are some small areas where limestone is at a depth of 40 to 60 inches and areas where the substratum is sand and gravel outwash. Runoff is medium. The erosion hazard is moderate. Erosion is a serious limitation. Most of the acreage is cropped to corn, small grain, or forage. Capability unit IIe-1; woodland group not classified; wildlife group 4.
Use and management of the
soils
Formation and classification of
the soils (factors of soil formation)
Additional facts about Columbia
County
13.3.3.) Shortcomings
Shortcomings in soil surveys are still the use of manual soil sampling techniques, which are costly and time consuming. Additionally, only point information with less support for fixed soil depths are collected. A deficit is definitely the lack of continuous data collection in the field. An example for a continuous soil data collection is ground penetrating radar, where radar waves penetrate into the soil and the signals are recorded. The signals can be used to identify different soil textures, dense layers or rocks (Doolittle, 1982). Profile cone penetrometers (PCP) are used to derive continuous measurements of penetration resistance, which can be related to soil properties such as soil texture, bulk density, and soil moisture. Typically, a probe is pushed into the ground from the surface and the cone index (the force per unit basal area) is measured by a data acquisition system. PCP were used by Campanella et al. (1981), Robertson (1990), and Lowery et al. (1994).
References
Campanella R.G., and P.K. Robertson. 1981. Applied Cone Research. In: Cone Penetration Testing and Experience. Norris R.M., and R.D. Holtz (eds.), ASCE, New York, 343-362.
Doolittle J.A. 1982. Characterizing Soil Map Units with the Ground Penetrating Radar. Soil Surv. Horiz. 23(4): 3-10.
Lowery B., and Schuler R.T. 1994. Duration and Effects of Compaction on Soil and Plant Growth in Wisconsin. Soil & Tillage Research, 29: 205-210.
Robertson, P.K. 1990. Soil Classification Using the Cone Penetration Test. Canadian Geotech. J. 27: 151-158.
back to:
[Main Page]
[Table of Contents]
[12) Soil Orders]
Soil maps and soil information systems are used for:
Land evaluations and tax
assessments
Farm and management recommendations
(e.g. fertilizer applications)
Prediction of erosion losses
Recommended conservation practices
Development of productive ratings
of soils
Soil potentials
Evaluations of sustainability in
land managment
Water quality evaluation (e.g.
nutrient leaching, pesticide yield)
Decision support systems
Water quality simulation models
(e.g. AGNPS, SWAT, GLEAMS)
Pedotransfer functions
Mined land reclamation
Planning, zoning, and other land
use conserns - local, state, and regional
Suitability of areas for septic
tank filters where the areas are not served by central sewage systems
Suitability for municipal sewage
effluent and sludge disposal
Highway route location
Building and real estate
development site loactions
Soil-related expert systems with
included simulation models (e.g. crop growth models)
many more........
Soil survey interpretation and soil information systems are prepared to help land users, planners, policy makers, legislative officials, engineers, and scientists to transfer technology about the use and management of soils - both agricultural and nonfarm - more accurately. The interpretations help predict potentials, limitations, problems, and managment needs for soils.
back to:
[Main Page]
[Table of Contents]
[12) Soil Orders]
Crisp soil maps have been developed during the last decades and are still the most widely used and available sources of soil information throughout most of the U.S. and the world. However, there are sophisticated techniques, which can be used to describe the gradually varying contiguity of soil properties. Most of them are research tools focusing on fine scales.
Most processes acting in soils (e.g. infiltration, drainage, weathering, and mineralization) may not produce crisp classes but a continuum in soil function and response. Using conventional soil maps in which the geographical distributions of soils are represented by homogeneous, sharply delineated polygons with attached information of soil attributes do not build up reality, because they ignore spatial variation in both, soil-forming processes and in the resulting soils. With crisp sets an individual is either a member or is not a member of a set (soil class). The corresponding membership function is 0 (false) or 1 (true). To derive a continuous soil classification fuzzy set theory (Zadeh, 1978) can be used. Observations are grouped into continuous classes in which individuals are assigned continuous class membership values instead of classifying the observations into exactly defined (hard or crisp) classes (McBratney et al., 1992; Burrough et al., 1992). With fuzzy sets the degree of membership can vary continuously between 0 and 1, hence soil classes can be expressed on an intermediate scale. Applications of a continuous soil classification using fuzzy set theory are given by Burrough (1989), Slater (1994), Irvin (1996), and McBratney et al. (1997).
Figure 13.5.1. Membership values for class a - d for a study site in Southern Wisconsin. The membership maps illustrate the degree to which high membership sites for each class are geographically clustered. The fuzzy classification has generated layer classes which are very similar to the horizons described in the county soil survey (Slater, 1994).
Figure 13.5.2. Cumulative membership values for exemplar profiles. The classes represent horizons described in the county soil survey; key attributes such as pH, carbon %, silt %, clay %, sand %, soil color, mottling were used to distinguish between horizons / classes (Slater, 1994).
References
Burrough P.A. 1989. Fuzzy Mathematic Methods for Soil Survey and Land Evaluation. J. Soil Sci. 40: 447-492.
Burrough P.A., R.A. MacMillian, and W. van Deusen. 1992. Fuzzy Classification Methods for Determining Land Suitability from Soil Profile Observations and Topography. J. Soil Sci. 43:193-210.
Irvin B.J. 1996: Spatial Information Tools for Delineating Landform Elements to Support Soi/Landscape analysis. PhD Dissertation, University of Wisconsin-Madison, 114 p.
McBratney A.B., and J.J. de Gruijter. 1992. A Continuum Approach to Soil Classification by Modified Fuzzy K-Means with Extragrades. J. Soil Sci. 43: 159-175.
McBratney A.B., and I.O.A. Odeh. 1997. Application of Fuzzy Sets in Soil Science: Fuzzy Logic, Fuzzy Measurements and Fuzzy Decisions. Geoderma, 77: 85-113.
Slater B.K. 1994: Continuous Classification and Visualization of Soil Layers: A Soil Landscape Model of Pleasant Valley, Wisconsin. PhD Dissertation, University of Wisconsin-Madison, 156 p.
Zadeh L.A. 1978. Fuzzy Sets as a Basis for Theory of Possibility. Fuzzy Sets Syst. 1: 3-28.
back to: [Main Page]
[Table of Contents]
[12) Soil Orders]
Spatial variability is governed by the processes of soil formation which are in turn interactively conditioned by lithology, climate, biology, and relief through geologic time. Spatial variability in soil systems belongs to two broad categories (Wilding et al., 1994):
Systematic (structured)
Random (unstructured and unknown
causes)
Systematic variability is a gradual or marked change in soil properties as a function of physiography, geomorphology and interactions of soil-forming factors. Systematic variation permits pedologists to partition spatial variability in soils of subsets of properties that constitute soil survey map units corresponding to geomorphic landscape elements (summit, shoulder, backslope, etc.). Large-scale spatial variability of a systematic nature may be as great or greater than long-range interval changes. An example of this are shrink-swell phenomena in soils that give rise to gilgai topographic relief variability in physical, and corresponding subsoil chemical and biological properties at intervals of meters or less. Fine scale variability occurs in aggregate ped units or microfabrics such as coatings of clay along void surfaces, zonation of oxyhydroxides, and concentrations of carbonates within the soil matrix. These distribution patterns reflect hydraulic flow, diffusion, immobilization and microbial colonization processes at micron and submicron scales in soil systems.
Causes of vertical and lateral anisotropy that yield spatial variability of a random nature over short-range or intermediate distances include: differential lithology, intensity of pedogenic weathering processes, hydrology, biological activity, erosion, deposition and pedoturbation; temporal effects of soil management; sampling and analytical errors. All of the above, except the latter two, may contribute to systematic variation, but the effects may be too subtle or complex to be discerned visibly or by measurement (Wilding et al., 1994).
The purpose of soil surveys is to partition spatial variability of landforms and soils. It is important to note, however, that appreciable variability still remains in mapping units of soil series (cartographic units) used to partition real geomorphic landscape components. Current NCSS standards for map unit composition require at least 75 % of the soils comprising the map unit to have similar interpretative ratings.
In Table 13.6.1 means and ranges in coefficient of variability (CV) are listed which have been reported in the literature for a selected number of soil properties sampled from equivalent horizons or depths within landsape mapping units of the same soil series (Upchurch et al., 1988). While these are only guidelines, they serve as useful indices in the absence of on-site data. The CV's for more stable properties range from 5 to 10 %, while for the more dynamic ones, they commonly range from 10 to 20 %, with extremes up to 35 %. Laboratory error analysis is property dependent but commonly with CV's less than 5 %. More permanent (stable) soil properties such as soil texture, mineralogy soil thickness, and color are less variable than temporal or more dynamic properties such as water content, hydraulic conductivity, redox state, salt content, biological activity, exchangeable cations and organic matter content. Properties which are measured and closely calibrated to a standard (e.g. texture, color, pH, etc.) are less variable than qualitatively accessed parameters such as soil structure, consistency, porosity, or root abundance. It should be stressed that the spatial variability in terms of pedological features may be significantly different from the spatial variability in terms of some functional feature.
Table 13.6.1. Relative variability of selected soil properties sampled within mapping units of a given soil series (Wilding et al., 1994).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Generally, the spatial variability in soils increases with the nature of the parent materials in the following order (Drees and Wilding, 1973):
loess 
glacial till
glacial outwash = glacial
lacustrine sediments = alluvium
pyroclastic and tectonic rocks
drastically disturbed materials
Elemental K = Ti Zr Fe Ca
No consistent trend is evident among A, B, and C horizons.
Figure 13.6.1. The relative variability of soil properties as a function of the permanence of the property (Wilding et al., 1994).
Figure 13.6.2. The relative variability of soil properties as a function of size of the sampling unit (Wilding et al., 1994).
Geostatistics can be used to analyze the spatial variability of soil attributes. Spatial continuity exists in most earth science data sets: Two data close to each other are more likely to have similar values than two data that are far apart. The regionalized variable concept is the basis for geostatistics, which states that a spatial variation of any variable might be expressed as the sum of three components:
a structural component, associated
with a constant mean value or a polynomial trend (deterministic
component)
a spatially correlated random
component (autocorrelative component)
a white noise or residual error
term that is spatially uncorrelated.
Regionalized variable theory is used to model the spatial dependence of soil properties by variogram analysis, which is required for kriging (spatial prediction). The variogram describes the degree of similarity between attribute values at sample sites x and x+h as function of their geographical separation or lag h. In variograms the distance between data points (x-axis) is plotted against the semivariance (y-axis). The semivariance is computed by the following equation:
Important to note is that variances as functions of the distance between measured points are considered, rather than the measurements of points. (Isaaks et al., 1989).
Figure 13.7.1. Example - Variogram: A spherical model describes the spatial varibility of the soil property. In this case there is no spatial dependence for points which are more than 75 m apart.
Different variogram models are used to describe the spatial relationship for different soil properties. The same soil property might show a different spatial variability in different landscapes. For example, clay content might show different spatial variability in a mountain landscape with steep slopes in contrast to an alluvial landscape with low slopes. Therefore, variograms cannot be transferred from one landscape to another without testing its validity.
Figure 13.7.2. Digital elevation model for a field on the West Madison Research Station, in Southern WI (Grunwald et al., 1998)
Figure 13.7.3. Cross-section showing the spatial distribution of cone index values (penetration resistance) from a summit position (elevation 329 m) to a lower landscape position (321 m). A variogram analysis and interpolation technique (ordinary kriging) was used (Grunwald et al, 1998).
Reference
Drees L.R., and L.P. Wilding. 1973. Elemental Variability within a Sampling Unit. Soil Sci. Soc. Am. Proc. 37: 82-87.
Grunwald S., K. McSweeney, B. Lowery, and D. Rooney. 1998. Continuous Description of Soil Attributes on a Landscape in Southern Wisconsin. Abstracts ASA-CSA-SSSA Annual Meeting, Baltimore, Maryland, Oct. 18-22, p. 253.
Isaaks E.H., and R.M. Srivastava. 1989. An Introduction to Applied Geostatistics. Oxford University Press, New York.
Upchurch D.R., L.P. Wilding, and J.L. Hatfield. 1988. Methods to evaluate spatial variability: 201-229. In: Wilding L.P., and J.L. Hatfield (eds.) - Reclamation of disturbed lands. CRC press, Boca Raton, FL.
Wilding L.P., J. Bouma, and D.W. Boss. 1994. Impact of Spatial Variability on Interpretive Modeling. In: Bryant R.B. and R.W. Arnold - Quantitative Modeling of Soil Forming Processes. SSSA Special Publ., No. 39: 61-75.
back to:
[Main Page]
[Table of Contents]
[12) Soil Orders]
Models are used to collect data in a systematic manner, describe systems or to predict relationships or behavior in order to extend our knowledge of soils. Models simply represent a series of approximations towards the 'truth'. Modeling is the classical method of science in operation. The model serves as the hypothesis; simulation and experiment serve to test it.
Soil landscape models provide descriptions of the gradually varying contiguity of soil properties in the landscape. A challenge in soil landscape modeling is to establish a relationship between soil and landforms and relate patterns to processes of pedogeomorphic evolution.
Characterization with respect to complexity (Huggett, 1975)
There are two approaches to build models of the soil system:
Isomorphic models describe
systems as a combination of all individual processes in the system
('a forest is described by describing each tree').
Homomorphic models develop
relationships for the system as a whole ('a forest is described as a
group of trees without describing each tree').
Isomorphic models are very complex and their implementation presents virtually huge problems; on the other hand, homomorphous models are often too general.
Characterization with respect to relative degree of computation, complexity, and level of organization (Hoosbeek and Bryant, 1993)
Hoosbeek and Bryant classify models using three characteristics: (i) degree of computation, (ii) complexity of model structure, and (iii) organizational hierarchy.
The first characteristic 'degree
of computation' distinguishes between qualitative and
quantitative models. Qualitative (conceptual) models comprise
mental, verbal, and descriptive models and are placed at one extreme
of the continuum. Mathematical models attempt to formalize these
abstractions as algorithms either as deterministic or as stochastic
approach (quantitative models). Deterministic models are based
on the premise that a particular set of characterization and input
information will produce one uniquely defined model prediction. By
contrast, stochastic models presuppose that input information and the
processes modeled can only be estimated within certain statistical
limits. Stochastic models use randomly selected distributed
parameters based on probability density functions.
The second characteristic
'complexity of model structure', distinguishes functional and
mechanistic models. Functional models use simplified or
empirical representations of basic process to accomplish their
predictions. This simplifies the model structure, reduces needed
input data, and reduces computational time. These models do not,
however, allow one to learn much about system behavior in terms of
basic process. These models generally depend either on statistical
relationships such as regression equations, or on simplified,
empirical conceptualization of water and chemical movement
characterized by static capacity terms, such as field capacity,
saturated water content, and bulk density. Alternatively,
mechanistic models incorporate fundamental mechanisms of the
processes involved. Their degree of complexity corresponds to the
model developer's concept of the present state of scientific
knowledge. These are generally deterministic models that recognize
that dynamics of the system depend on rate processes, such as the
rate of water flow, or the kinetics of mineral dissolution. These
models are more complex because they require the use of differential
equations and iterative procedures to provide solutions.
The third distinction is based on
the 'organizational hierarchy', which describes at which level
a model aims to simulate a natural system. Each level can be regarded
as a system by itself, and can be seen as a combination of subsystems
at lower levels or as a subsystem of higher levels. Each level
integrates the knowledge of subsystems at lower levels, which means
that investigations at a subsystem level, e.g., i-1, provide a
mechanistic understanding of a model at the i-level. In case of
pedogenesis, the pedon is placed at the central i-level in the
hierarchy. Investigations of process and variability must be
investigated at a level of resolution appropriate for the level at
which a model aims to simulate the system.
Figure 13.7.1.1 Classification of models (after Hoosbeek & Bryant, 1993).
Table 13.7.1.1. Scale and associated systems (Wagenet et al., 1994).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Pedologists have developed qualitative landscape models based on the state-factor analysis approach. The five soil-forming factors suggested by Jenny (1941) are: (i) climate, (ii) organisms, (iii) topography, (iv) parent material, and (v) time. This soil landscape model formalizes the relationship between soil, landform, and processes of the landscape resulting from soil genesis. Formalization of the factors of soil formation, the roots of which can be traced back to Dokuchaev, fostered the concepts of soils as integrated components of ecosystems. Factorial models explain system characteristics in terms of external variables but reveal little about details of soil-system dynamics. Relationships derived from climo-, bio-, topo-, litho-, and chronofunctions are common in pedologic literature; however, to date, no one has developed a mathematically rigorous quantitative solution to any univariant state-factor equation. One problem with univariant state-factor equations is the partial dependence and frequent interactions among the factors. State-factor analysis provides a useful, homomorphic, conceptual model of soil systems; however, factorial approaches yield little for enumerating and assimilating the processes active in soil systems. Correlations between soils and state factors, even when clearly shown to be causative relationships, do not in themselves explain any of the mechanisms. Furthermore, such conceptual models are extremely difficult if not impossible to test and validate. However, the factorial model has contributed to our understanding of soil formation.
Runge (1973) has proposed a factorial model, with priority to the factors (i) water available for leaching, (ii) organic matter production, and (iii) time. Organic matter production can be correlated with parent material (source of nutrients) and vegetation and water available for leaching with climate and relief in Jenny's model. However, Runge's model places more emphasis on processes active in soil systems rather than environmental variables external to the soil and he stresses energy fluxes to the soil system. Water available for leaching is essentially an effective agent for utilizing gravitational energy whereas organic matter production is an expression of radiant energy. Conceptually, Runge suggests that soil formation is analogous to energy fluxes acting on a chromatographic column.
Milne's (1935) pioneering work in East Africa spawned the catena concept, which emphasized integrated relationships among slope/landform and hydrology/morphology, though with a strong geological bias.
Huggett (1975) describes a soil-landscape system driven by a constant flux of material and energy between the system and its environment. He defines 'soil-landscape systems' as the basic 3-D functional unit of soil systems. Any functional unit must include movements tangential to the surface as well as vertical movements. The model can accomodate geomorphic processes just as readily as pedogenic processes. Huggest considers the soil-landscape consist of three materials: (i) soluble, (ii) plasmic, and (iii) skeletal. These materials are transmitted through and transformed within the system. Plasmic components are basically affected by pedogenesis whereas the skeletal components are affected by geomorphogenesis. Huggett's view of the soil-landscape system functions is that of an open system, however, a black box. Therefore, the model contributes little to our understanding of processes operative within the system. Nevertheless, Huggett indicates that soil-landscape systems exhibit all the features of a dynamic system model: system components, boundaries, interrelationships, transport and conservation of material and energy, and time flow.
Almost all soil landscape models recognize the importance of hydrology to pedogenic and geomorphic processes. Since water movement is considered to be one of the most important driving forces in soil genesis and landscape evolution, the location of flow paths and water distribution on slopes is important for predicting pedogenic variability. Topography affects parent material composition through deposition of loess and eroded material, soil moisture by surface and subsurface flow, and soil temperature by solar radiation. These soil landscape models hypothesize that there is a close relationship between soil attributes and landscape position. Examples are given by Pennock et al. (1987) and Irvin (1996).
The system approach which is process oriented is concerned with fluxes of material and energy through soil systems and thus relates driving forces of pedogenesis to soil-system dynamics. Simonson (1959) presented one of the earliest attempts to model soil formation using a system or process-response approach. The model suggests a generalized scheme for organizing and studying processes operative within soil systems. According to Simonson, soil genesis consists of two steps: (i) parent material accumulation; and (ii) differentiation of horizons in the profile which is a function of additions, removals, transfers and transformations within soil systems. These 4 processes are very general in order to cover the entire range of specific processes active in soil genesis. Nevertheless, the general processes do provide a framework for conceptually organizing processes to facilitate understanding. Simonson postulates that all processes are proceeding simultaneously in all soils and that only rates differ among soils. The ultimate nature of soils is governed by the balance among all the processes. Such a model used an isomorphic approach and faces the problem to identify the magnitude and complexity of all processes occuring in soils.
The development of quantitative pedogenic process models has gained interest in the last decade due to a better understanding of pedogenic and geomorphic processes, significant advances in computer power, geostatistics, and available data in soil information systems. Models of pedogenesis based on physical, biological, and chemical processes are isomorphic in character. Processes that lead to soil formation are a combination of weathering reactions and biological activity on a particular parent material and topography, under given climatic and anthropogenic circumstances, with the products redistributed through soil profile by water. Continuation of such processes over time and space results in the differentiation of soils in both the horizontal and vertical dimensions. Nonetheless, due to the complexity of natural systems, quantitative mechanistic models often do not exist. However, they are superior over conceptual models because they are fully transferable to other sites and landscapes. Quantitative pedogenic process models are extremely difficult to validate because of the long time range of pedogenic processes. Most quantitative soil models developed to date focus on discrete chemical, physical, or mineralogical processes and isolate these meachanisms from other processes occuring simultaneously in the soil. Many of the quantitative models focus solely on a single system (e.g. leaching of chemicals through the soil profile, physics of heat and water transport through the soil). The requirements for quantitative pedogenic process modeling are: (i) mathematical description of interactions of the pedosphere with other geospheres, (ii) description of the complex mechanisms of hydraulic, pedogenic and geomorphic processes, (iii) supply of input variables (soil attributes, which often show great spatial variability). Examples for pedogenic process models are presented by Suarez et al. (1994) and Hoosbeek et al. (1994).
Gerrard (1990) demonstrated that not all slopes possess fully integrated soil systems. Different parts of the slopes may act independently. Such discordance is likely in landscapes composed of a variety of parent materials that have been subjected to climate, land use, and pedogeomorphic change. An important question that needs to be considered in development of soil landscape models is: How strong are the relationships among soils and landforms across the landscape and at what scales are the relationships most evident?
No single model can accomodate observed variability operating at the local scale (Gerrard, 1990). Conceptual models appear limited to unraveling the 3-D complexity inherent in soil-landform relationships. Commonly methods using a 1- or 2-dimensional design are used to address a 3-D problem, i.e., to describe the distribution of soils and soil attributes in 3 dimensions. On the other hand quantitative pedogenic process models are often too complex. They tend to isolate phenomena and describe every detail of the soil system becoming greater and greater in detail.
Further Reading
McSweeney K., P.E. Gessler, B.K. Slater, G.W. Petersen, R.D. Hammer, J.C. Bell. 1994. Towards a New Framework for Modeling the Soil-Landscape Continuum. Factors of Soil Formation: A Fiftieth Anniversary Retrospective. SSSA Special Publication 33: 127-143.
Wilding L.P., N.E. Smeck, and G.F. Hall. 1983. Pedogenesis and Soil Taxonomy I. Concepts and Interactions: 66-81.
References
Butler B.E. 1959. Periodic phenomena in landscapes as a basis for soil studies. Soil Publ. 14, CSIRO, Div. of Soils, Canberra, Australia.
Gerrard A.J. 1990. Soil variations on hillslopes in humid temperate climates. Geomorphology, 3: 225-244.
Hoosbeek M.R., and R.B. Bryant. 1993. Towards the Quantitative Modeling of Pedogenesis - A Review. Geoderma 55: 183 - 210.
Hoosbeek M.R., and R.B. Bryant. 1994. Developing and Adapting Soil Process Submodels for Use in the Pedodynamic Orthod Model. In: Bryant R.B. and R.W. Arnold - Quantitative Modeling of Soil Forming Processes. SSSA Spec. Publ. 39: 111- 128.
Huggett R.J., 1975. Soil Landscape Systems: A Model of Soil Genesis. Geoderma 13: 1-22.
Irvin B.J. 1996: Spatial Information Tools for Delineating Landform Elements to Support Soil/Landscape Analysis. PhD Thesis, University of Wisconsin-Madison.
Jenny H. 1941. Factors of Soil Formation. McGraw-Hill, New York, N.Y.
Milne G. 1935.composite units for the mapping of complex soil associations. Trans. 3rd Int. Congress Soil Sci, 1: 345-347.
Pennock D.J., B.J. Zebarth, and E. de Jong. 1987. Landform Classification and Soil Distribution in Hummocky Terrain, Saskatchewan, Canada. Geoderma, 40: 297-315.
Runge E.C.A. 1973. Soil development sequences and energy model. Soil Sci., 115: 183-193.
Simonson R.W. 1959. Outline of a generalized theory of soil genesis. Soil Sci. Soc. Am. Proc., 23: 152-156.
Suarez D.L., and S. Goldberg. 1994. Modeling Soil Solution, Mineral Formation and Weathering. In: Bryant R.B. and R.W. Arnold - Quantitative Modeling of Soil Forming Processes. SSSA Spec. Publ. 39: 37 - 60.
Wagenet R.J., J.L. Hutson, and J. Bouma. 1994. modeling Water and Chemical Fluxes as Driving Forces of Pedogenesis. In: Bryant R.B. and R.W. Arnold - Quantitative Modeling of Soil Forming Processes. SSSA Spec. Publ. 39: 17 - 35.
back to: [Main Page]
[Table of Contents]
[12) Soil Orders]
Samples can be taken with a bucket auger and analyzed in the field (e.g. soil color, soil structure) or in the laboratory (e.g. Fe content, bulk density). There are two different methods for sampling: (i) sampling at fixed depths (e.g. 30, 60, 90-cm depth), or (ii) sampling in each horizon. Conventionally, soil sampling is carried out using different sampling designs. Common sampling designs are:
Grid sampling: A
grid with suitable spacing is placed on a landscape to be studied.
Sites can be selected at intersections of the grid lines or within
the grid cells. Grid sampling does provide equally spaced
observations and it reveals any systematic variation across the tract
under study. The drawback in geostatistical analysis is the equal
distance between all sampling points. It should be noted that there
is no randomization accociated with grid sampling, therefore, the
assumptions underlying several statistical analysis (e.g. ANOVA -
analysis of variance) can not be fulfilled.
Random sampling:
Sample locations are selected at random, with equal probabilities of
selection and independently from each other. The rationale is to
exclude any form of bias, such as a conscious or even unconscious
process of discriminatory selection on parts of the individuals. The
technique has advantages of being statistically sound and unbiased,
however, random samplings tend to cluster spatially (nonuniform
density of observations per unit area and of dispersion of sites over
the delineations) and are not likely to detect and measure systematic
variation.
Random stratified sampling: The area is first divided into a number of
sub-regions, called strata, and then random sampling is applied to
each of the strata separately. The sample sizes in the strata may be
chosen such that the probabilities of the locations of being sampled
differ between strata.
Transects: Soil
samples are taken along straight lines across a landscape. The
spacing between sampling points might be equal, nested, or random.
Transect sampling reveals spatial variability along a line (often
downhills), however, spatial variability in other directions is
neglected.
Target sampling: Two
or more attributes (e.g. topographic attributes such as slope,
aspect, plan or profile curvature) are used to identify homogeneous
and heterogeneous patterns. The goal is to identify 'representative
sampling points'. This is a technique which minimizes the
effort (costs) and maximizes the information content, on the
assumption, that the sampling points are representative for the total
data set (study area). It should be noted that
there is no randomization accociated with target sampling, therefore,
the assumptions underlying several statistical analysis (e.g. ANOVA -
analysis of variance) can not be fulfilled.
Different sampling approaches must be used depending on the objectives, which are strongly influenced by scale. Each experimental design has constraints and strengths with regard to the analysis of data.
back to: [Main Page]
[Table of Contents]
[12) Soil Orders]